达梦数据库、MySQL MERGE INTO 用法

1. 需求描述

在我们日常开发中，经常会有些数据是唯一的，比如人员的信息、车辆基本信息等等，我们在数据库中只存一条数据，我们在批量导入数据的时候，有重复数据是很难受的事情，我们用MERGE INTO 就可以解决这件事情

2. MERGE INTO 语句简介

使用 MERGE INTO 语法可合并 UPDATE 和 INSERT 语句。通过 MERGE 语句，根据一张表（或视图）的连接条件对另外一张表（或视图）进行查询，连接条件匹配上的进行 UPDATE（可能含有 DELETE），无法匹配的执行 INSERT。其中，数据表包括：普通表、分区表、加密表、压缩表和堆表。

语法格式

MERGE INTO <merge_into_obj> [<表别名>] USING <表引用> ON (<条件判断表达式>)
<[<merge_update_clause>] [<merge_insert_clause>]>
<merge_into_obj> ::= <单表引用> | <子查询>
<单表引用> ::= [<模式名>.]<基表或视图名>
<子查询> ::= (<查询表达式>)
<merge_update_clause>::=WHEN MATCHED THEN UPDATE SET <set_value_list> <where_clause_null> [DELETE <where_clause_null>]
<merge_insert_clause>::=WHEN NOT MATCHED THEN INSERT [<full_column_list>] VALUES <ins_value_list> <where_clause_null>;
<表引用>::=<普通表> | <连接表>  详见《DM8_SQL语言使用手册》第四章 数据查询语句
<set_value_list> ::= <列名>=<值表达式| DEFAULT> {,<列名>=<值表达式| DEFAULT>}
<where_clause_null> ::= [WHERE <条件表达式>]
<full_column_list>::= (<列名>{,<列名>})
<ins_value_list>::= (<插入值>{,<插入值>})

3. 具体SQL语句

*Mapper.java

void mergeSaveCarInfoList(@Param("list") List<TbCarInfo> list);

*Mapper.xml

 <insert id="mergeSaveCarInfoList" parameterType="com.shxp.project.target.domain.po.TbCarInfo">
        MERGE INTO tb_car_info t
        USING
        (
        SELECT
        car_no,car_type,car_color,car_normal,car_company,created_by,created_date,owner,id_card,home,tel,remark,update_by,update_date,created_by_id,update_by_id,flag,car_type_name,car_color_name
        FROM (
        <foreach collection="list" item="item" separator="union all">
            SELECT #{item.carNo,jdbcType=VARCHAR} as car_no,
            #{item.carType,jdbcType=VARCHAR} as car_type,
            #{item.carColor,jdbcType=VARCHAR} as car_color,
            #{item.carNormal,jdbcType=VARCHAR} as car_normal,
            #{item.carCompany,jdbcType=VARCHAR} as car_company,
            #{item.createdBy,jdbcType=VARCHAR} as created_by,
            #{item.createdDate,jdbcType=VARCHAR} as created_date,
            #{item.owner,jdbcType=VARCHAR} as owner,
            #{item.idCard,jdbcType=VARCHAR} as id_card,
            #{item.home,jdbcType=VARCHAR} as home,
            #{item.tel,jdbcType=VARCHAR} as tel,
            #{item.remark,jdbcType=VARCHAR} as remark,
            #{item.updateBy,jdbcType=VARCHAR} as update_by,
            #{item.updateDate,jdbcType=VARCHAR} as update_date,
            #{item.createdById,jdbcType=VARCHAR} as created_by_id,
            #{item.updateById,jdbcType=VARCHAR} as update_by_id,
            #{item.flag,jdbcType=VARCHAR} as flag,
            #{item.carTypeName,jdbcType=VARCHAR} as car_type_name,
            #{item.carColorName,jdbcType=VARCHAR} as car_color_name
            FROM dual
        </foreach>
        ) tmp
        ) a
        ON (a.car_no = t.car_no)
        WHEN matched THEN
        UPDATE SET
        t.car_type = a.car_type,
        t.car_color =a.car_color,
        t.car_normal= a.car_normal,
        t.car_company = a.car_company,
        t.created_by = a.created_by,
        t.created_date = a.created_date,
        t.owner = a.owner,
        t.id_card =a.id_card,
        t.home =a.home,
        t.tel = a.tel,
        t.remark = a.remark,
        t.update_by = a.update_by,
        t.update_date = a.update_date,
        t.created_by_id = a.created_by_id,
        t.update_by_id = a.update_by_id,
        t.flag =a.flag,
        t.car_type_name =a.car_type_name,
        t.car_color_name =a.car_color_name,
        t.is_delete ='0'
        WHEN NOT matched THEN
        INSERT
        (car_no,car_type,car_color,car_normal,car_company,created_by,created_date,owner,id_card,home,tel,remark,update_by,update_date,created_by_id,update_by_id,flag,car_type_name,car_color_name)
        VALUES
        (a.car_no,a.car_type,a.car_color,a.car_normal,a.car_company,a.created_by,a.created_date,a.owner,a.id_card,a.home,a.tel,a.remark,a.update_by,a.update_date,a.created_by_id,a.update_by_id,a.flag,a.car_type_name,a.car_color_name)
    </insert>

4. 应用场景

xlsx数据导入